Mobile devices are currently generating mobile (spatio-temporal) data streams at a high-rate for sending to backend servers in real time. Especially with the rise of sensor technology, there are a plethora of mobile devices generating such high-rate mobile data streams. While the mobile data streams are critical sources of mobility analytics supporting advanced mobile services, they also raise new challenges of parallel processing for trajectory-featured data streams.
A dispatching service is a component that supports real-time message processing. With e-partitioned areas and zones, a dispatching service can enable high quality dispatching of mobile data messages according to the geo-partition information from a partition service. In location based message dispatching, longitude and latitude information of message is generally used.
Internet Protocol based dispatching may not be feasible for message dispatching high speed movement objects. For instance, a technique that dispatches by region shape may involve computation of complex relationships between longitude and latitude. Geometry computation can be costly, and in the presence of large clusters of nodes, the dispatching by region shape may become very inefficient. A movement context switch also may pose challenges. For instance, when a moving object moves toward a boundary and prepares to cross the boundary, many application scenario may exist that require switching context to prepare for roaming such as key switch, register information switch, context table switch and refresh, all of which information may generate additional burden for computation and dispatch in an attempt for understanding the distance between the moving object and the boundary. Such factors in dispatching messages from and to moving objects contribute to bottleneck in the overall communication and/or messaging system.